Inside Pontiac’s CTV-First DSP: Keith Gooberman Built the Only Architecture for CTV

CTV advertising is exploding and this is the Adotat show. While the legacy DSPs [music] keep lumbering around like they're still hunting cookies in 2014, Keith Gubberman strolls in with Pontiac

[music] and rewires the whole playbook.

Leaner, faster, actually built for CTV instead of pretending.

>> If [music] the DSP world needed a plot twist, this is the guy delivering it.

>> Hi, I'm Pesock Latin and this is the W show. I am here with Keith Gooberman of

show. I am here with Keith Gooberman of Pontiac. Legacy DSPs like the Trade

Pontiac. Legacy DSPs like the Trade Desk, Google, DB3, Amazon DSP dominate, but they were built for cookies and banner ads, not CTV. Pontiac claims to

be a leaner, cheaper, and actually designed for programmic 2.0. First, most

important question of the day. How are

you Keith?

>> Doing well, Pesock, how are you doing?

>> Thank God. Doing great. Let's just cut through the fluff and get to the point.

Does the industry actually need another DSP in 2025?

>> That's a great question. multi-billion

ver you know some people use the t-word with trillion the advertising space it's it's going to continue as Jeff Green says in all of his his uh public statements during during quarterly

announcements it's going to continue to move to the programmatic pipes it's going to move towards more automation and done through the programmatic

auction as as we know it and love it so when you're talking about a billion dollar industry in just the fees component, okay? Like hundreds of

component, okay? Like hundreds of millions of dollars made by some of those companies you just mentioned, there's definitely room. There's room

for competition. There's room for people with new ideas doing things differently.

There's always room. It has dried up in terms of raising capital to build a DSP.

That's why my story is a lot about how I kind of bootstrapped it off of programmatic mechanics success, but it is hard to raise money for because the model is difficult to get right and to make a difference. you know, you as you

said, as we've talked about a little bit, we don't have cookies. Okay, that's

a big difference. That's a a very large difference in terms of our infrastructure costs, the way we think about it, the way our DSP works um compared to others. Uh how much room is

there for competition? I'm not sure. Do

I think that there's room for us? I

think so. But again, our model is not we're not taking the same model from yesterday and just kind of saying, "Hey, we have a new name and more marketing dollars and more yogurt for you if you come to our happy hour."

>> If last touch attribution [music] worked, your waiter would get full credit for the meal. We make sure the chef gets credit, [music] too.

Incremental.

I want to ask you, so when you say the incumbents are light battleships, and I'm going to go with this metaphor a little bit. Is there an example when a

little bit. Is there an example when a client came to you was actually sinking under the weight and did you have to hand out life jackets?

>> Yeah. So, we kind of equate the the older model with cookies to the battleship thing because cookies so as a DSP sees a first impression from an SSP.

This is the way they've been built for a very long time. As they see a first impression, they don't bid. They simply

use that as an opportunity to place the cookie ID from the SSP into a match table so that they know who the user is.

Then the second time they bid on it. And

the the number of cookies and the cookie maps that these DSPs have to maintain are enormous. And because you need to

are enormous. And because you need to look up what segments each user falls into before the bidder can even respond.

like the bidder can't respond until it knows what segments the user's in so that it knows what lines are going to bid on it. Those lookups of what segments those is a tremendous amount of

computing power. Every bid costs the DSP

computing power. Every bid costs the DSP a sizable amount of computing power. Now

at scale it all works out fine as we know. Um but it's still very complicated

know. Um but it's still very complicated and very cumbersome to do it that way.

And um the amount of servers that you need the amount of the amount of servers you need to see 30 million ads a second so that your remarketing works well is

expensive. You need to have a very large

expensive. You need to have a very large number of computers running at all times across the world that can see all this traffic and these things make the maintenance of the infrastructure the

core bidder very expensive. But if you remove the cookies and you're just bidding on deals that are sending traffic in however they've shaped it, the bidder, you know, everybody when you use the term bidder, I think everybody

thinks of a little person with a paddle, right? Which obviously it's all

right? Which obviously it's all automated, but in a way our bidder is the paddle thing. It's saying, listen, the SSP is going to send me all the traffic I want. They've got we've got a sports deal targeted to greater

Minneapolis. That's what I want. I want

Minneapolis. That's what I want. I want

it running from 6:00 p.m. to 10 p.m. and

I want it. That's what I want. The SSP

can send that, but you still need a DSP to upload the creatives to pace to to be able to buy maybe some optimization. So,

with curation moving the tech more to the sell side, we have seen clients be very receptive to our lighter DSP model.

And I would say that they are because our model allows us to be thinner margin. We are very prom we promote and

margin. We are very prom we promote and we talk about and we're very friendly with the publishers because again we're not like we're taking less of the dollar than the next guy. So they're like they're like yeah this DSP is better

than that one because we make more of our money. So we align very well. The

our money. So we align very well. The

way we talk about it is much more about getting to a publisher. It's not

necessarily about our own tech and all that although we've built a lot of things around measurement and all that.

I don't know if that was a good >> So hypothetical question. at the trade desk was CocaCola. Is Pontiac the artisal kombucha, the offbrand RC cola, or that sketchy Chinese unlabel you can

find at the bodega fridge? I'm going to go that we're a little bit more of the flavored seltzer water of today, which is the DSP of yesterday, the Coca-Cola,

the full soda, you know, a staple for many years, enjoyable. But for those that want the carbonation, a little less calories, a little lighter, I think that we'd be more of that because

>> So, no following like RC Cola.

>> Yeah. Well, the re, you know, we have to to go in there and win brands that spend larger dollars and care about saving points on the tech fee, we better have

all the top CTV channels. So to get the Disney, NBC and all these things, serious radio, all these all these, you know, not walled gardens, but fortified

publishers to get in, we have to be super legit, super clean. We can't be doing anything trying to steal audience data, like, god forbid, you know, we have to be very legitimate. We had to

build a full-blown soda brand and bring it to market. But I would say that we're trying to do it a a little The flavor is different. It's not another soda. It's a

different. It's not another soda. It's a

It's sheltered water for those that that want us.

>> Eric Troutman, the gazar of the TCPA, isn't just the guy you call when your phone rings too much. He's the guy you must call. Need to navigate the legal

must call. Need to navigate the legal minefield of rooc calls, texts, and telemarketing nightmares.

Troutman doesn't just give you advice.

He hands you a road map, a compass, and a do not press this button sign.

Practical? Yes. Actionable? Absolutely.

And if you're not calling him first, let's be honest, you're probably making a mistake.

>> If I just put this clip out there like that, no one's going to have any idea what we're talking about.

>> That's okay.

>> So, can you walk me through a campaign where Pontiac outperformed a giant? And

I want to know what cocktail you then celebrated with. And you can't say

celebrated with. And you can't say Michelob Alter. I'm I'm just shutting

Michelob Alter. I'm I'm just shutting everything down.

>> Hey, what's my current drink of choice for for a victory? Is that what the question was? I didn't hear you. Well,

question was? I didn't hear you. Well,

it was like it was more like what campaign made you go into the victory dance and drinking.

>> Yeah. Oh, yeah. So, um, so in Pontiac, when you're setting up lines, a lot more of our focus is on the PMP kind of by line. Okay. So, a typical CTV campaign

line. Okay. So, a typical CTV campaign would have a couple of lines dedicated, you know, a Disney line, an NBC line, a a Roku line, the the big guys, a pho line, some of the aggregators, and then

it might have some more performance kind of like more of your remnant like inventory that you can get from some of the fast channels. Okay, so you have a a campaign, it's got 11 lines. Okay, now

what they you what they do being the users or you know, people using the platform, some of these come through with a lot of content signals. Okay, so

Disney famously doesn't let you see the content until you see the genre. Okay,

so you could do some genre targeting on that first Disney line for some of the fast channels or some of the aggregators or the VMVPDs, which is how like Phoes themselves, kind of like a modern cable

company, aggregating content and distributing it. Those guys pass more

distributing it. Those guys pass more content signals. They might show you the

content signals. They might show you the channel they're watching in the Pho app.

You might know that they're watching cooking versus uh home improvement versus a sports versus a talk show. And

some of them even pass the showlevel data. Okay, so some of the cable

data. Okay, so some of the cable companies with their carrying rights have the right to show you exactly what somebody's watching. The actual episode

somebody's watching. The actual episode of Law and Order SVU. It's hard to say how valuable that is exactly, but again, there's different ways to use this for different buyers. 11 lines all set up

different buyers. 11 lines all set up with different content targeting based on the publisher using our pixels. So we

have our own IPbased pixel that will basically do attribution across these different channels across these different content pieces. And so you can

set the platform to optimize to web-based conversions on the CTV spend.

And what it does is it focuses on the word allocation, which not that that's a foreign word to anybody in this business, but it's not typically the way DSPs talk about optimization. They talk

about it as bid changes and machine learning kind of like how we orchestrate all these bids.

>> So you're going back into media buying.

You is this media buying 101 from >> You know what? You know what? Again,

we're talking about allocating budgets.

We're talking about moving budgets around based on performance, but in a very, you know, you nailed it. A classic

way. And again, you could ask the platform to auto allocate, but other DSPs that are built for display, they want to get the CPMs down. They're like

designed to like lower the CPM until they don't win enough inventory. That

doesn't really work on CTV, okay? Your

floor price with these really nice providers. That's not where you're going

providers. That's not where you're going to get your optimizations. If you want to change the CPMs, you got to allocate money into different buckets. You're not

just like bidding less across the mall.

The platform is designed in I think you nailed it like in a in a way that a more you know a media buyer 101 would kind of think about it and the platform reflects that and that's you could do this across

six curated deals that are all curating different display. So it's not only for

different display. So it's not only for CTV but again it's the word allocation is much is much uh more prominent for us.

>> So I want to move on. I want to talk a little about you being bootstrapped versus VC pumped behemoths. You didn't

take a dime of VC cash. To me, that feels a little bit like showing up to a Formula 1 track in a Toyota Corolla, which is my wife's car, but you're lapping the Ferraris. So, why did you go

bootstrapped? Was it bravery,

bootstrapped? Was it bravery, stubbornness, or you just had no money?

>> Okay. So, I wish I could say that it was some fantastic foresight. But what we did was we built a managed service business primarily over our friends at

the time called App Nexus and then called Nander and now called the Microsoft something still not totally but it's something it's in there.

They're shutting down the BSP by the way. I'm sure you're aware of that. So

way. I'm sure you're aware of that. So

we built a managed service business. I

then brought on our CTO. Our CTO is basically a friend at this point but it's a business colleague. But Eric

Thorson is an exceptionally gifted computer scientist. and I brought them

computer scientist. and I brought them on and we just started building. So we

first built tools over the app nexus buyside API to help us with optimization and allocating money across. We want a lot of very small

across. We want a lot of very small clients, modest budgets. So how to manage campaigns that only spend $100, you know, every two days, three days is different than managing a campaign that spends 5,000. So we built tools to do

spends 5,000. So we built tools to do that. We then built a front-end version

that. We then built a front-end version of a DSP. So Pontiac was born as a front-end version of a DSP that just used the API into Xan App Nexus and Xander some somewhere in here they

changed to Xander and we built an integration to Stripe the credit card company and we were very close with the app nexus people. So, I kept saying to them, Xander people now, hey, are you getting leads that are people that want

to test it for $5,000? Like, you're not signing an MSA for that. Why don't you send me those? I have a self-service platform now that they could swipe a credit card and use it and go from there. And that business started to

there. And that business started to work. And they started sending clients

work. And they started sending clients and we started onboarding clients and the business started to work. They were

smaller so they didn't mind the upcharge that I had to do to justify my own fee before the App Nexus fee, Xander fee.

So, it all worked out fine. Works worked

great. platform is still live by the way so people still use this but that started to work and at that time I because that started to work and we were successful in building that business now we have two business program mechanics

and Pontiac because we started building the Pontiac piece we invested in engineering talent okay so we were able to hire engineers we got a couple of very great engineers we got a a

brilliant PhD out of Poland he's actually a rock drummer this guy Maric is also amazing and a team and at that point the platform was and just an API over app. That's just we were kind of

over app. That's just we were kind of done. Like it didn't need much more

done. Like it didn't need much more building. It was running. It was going

building. It was running. It was going fine. CTV comes along and I'm going,

fine. CTV comes along and I'm going, "Guys, we're paying the regular DSP fees, but between you and me, they're sending deals. It can't be more than

sending deals. It can't be more than 2,000 QPS." Again, my example, only live

2,000 QPS." Again, my example, only live sports, only from in Minneapolis, like the actual deal itself was like many less QPS. So then with the engineers,

less QPS. So then with the engineers, the company working, I had the engineering talent. We had a front end

engineering talent. We had a front end already built in 20 end in 2019, beginning of 2020. I said, "Guys, we should build a bidder." Like, but I'm not going to build it. You would you build a bidder? And they were like,

"It's going to take us about two years."

And I was like, "Great." Now, this goes a little bit to the VC thing. I said,

"Great, sure, let's go start building."

That is a long time in the VC world to just say, "Right, >> go build something." I didn't even have to ask anybody. I was just like, "I like that. Why don't we do that?" And so that

that. Why don't we do that?" And so that speed, that decision-making, that ability to kind of run fast, they built it. So now it's 2022. They had to get

it. So now it's 2022. They had to get integrated with Magnite and Freewheel and Index and Xander Monetize. That took

another year and approvals the the contract with the mouse on it, the Disney one and those. It takes time. But

then in 2023, we really went to market and we had ourselves a buying a a platform done for buying. Then curate

took off and like now we have a very efficient tool to buy against curated deals. So, oh that's cool.

deals. So, oh that's cool.

>> Do you consider yourself a curator?

>> No, no, no. But if you're a curator, not you, but if the proverbial use is a curator, >> I curate some stuff as you say.

>> There you go. If you then push deals to DSPs, you know, the DSPs again aren't doing that much. They just want to do pacing and the creative and make sure that it runs and stops when it's supposed to stop. Now, listen, we have

unbelievably advanced ways to do day parting so that it mimics TV schedules.

We have custom algorithms so you could literally target those shows and mimic TV like your TV schedule or off your schedule or whatever it is. So it's not to say that there's not advanced tools to do buying but we had no cookies and

it really with the curation thing taking off more people are doing decisioning on the sell side and they want the DSP to be lighter. So that that timing, all

be lighter. So that that timing, all that working and then again as I said for for the scaled buyers, we are several percentage points thinner than the other DSPs on the market. And that

type of that type of revenue model is not about making money today. Okay? If

you take a single digit percentage, you better you bet you're not making money today. You might get a big client to

today. You might get a big client to sign up and then you might get them to scale and at some point two years down the road it'll start to make some actual

money. And that is an equation I again

money. And that is an equation I again we didn't set out to get here because again it was kind of this way CTV let's build. But now that we're here it's very

build. But now that we're here it's very hard to get a bidder to scale to work to work worldwide to have all these features and to do it at such a thin

margin. If I went into a VC room and I

margin. If I went into a VC room and I told them the plan, they would tell me it's it's basically not going to work.

And they'd be right unless we use the power of time and a couple of balls bounced our way. And that I think that's what's that's what's happened now.

You've said before that bootstrap makes you nimble, but doesn't it also make you a bit irrelevant on scale? Aren't you

screaming into the wind while Amazon builds a Death Star? I mean, honestly, does Death Bezos or Darth Bezos care about indie DSPs? Wonderful question and a very difficult thing to pontificate

when you're on the outside of those walls. Okay, I'm not sure. I believe

walls. Okay, I'm not sure. I believe

that publishers and adtech companies are a relatively large client set on AWS. I

could be wrong, but outside of healthcare, advertising records and advertising information and advertising machines need a lot of machines. So,

like he's they're definitely coming after the ad spend via the DSPs. And

would having a lot more dollars for marketing make it easier? I believe that that would be nice. I would I I'm not going to stand here and say that I wouldn't like a couple million dollars to to do some events and uh things and

and we have to keep it a little bit slimmer on that side.

>> Is the end game selling to Disney?

>> What >> is the endgame selling to Disney?

>> The end game is to build a DSP that is scaled, competitive, a true value ad for clients that use it and then see. I

don't have any exit plans and I'm not looking for them. And but as going back to your last point for one second, the the idea of does bootstrapping hurt in that way. You know, it can it can in the

that way. You know, it can it can in the way that some people want uh a lot of flash, a lot of flash. They want a lot of noise around it. Thanks to you and doing this with me today and a few other little interviews here and there, we

definitely have gotten our name out there. And listen, I go back a little

there. And listen, I go back a little bit to the engineering talent we have again, Eric and Mark and and a few others and the way we've done it and the way we've designed the platform. We've

used a lot of platforms. We know a lot about buying ads. The DSP is pretty nice to use if you use a DSP. Um, no crazy artistic tables getting in your way or whatever whatever other things are

offering. And you know what? I think

offering. And you know what? I think

we're doing pretty good on the product itself. You know, you got to use it. You

itself. You know, you got to use it. You

got to say, I'm going to give it a shot.

But when we get people using it, they see that it's better control, better results, better transparency into what they're buying and better uh pricing, all these things. And listen, I'm not

I'm not here I'm not one of those companies that goes with total flash. If

you're going to try Pontiac, you're going to be a serious buyer. You're

going to be looking to make your buying more efficient. You're going to be

more efficient. You're going to be excited about the tools we built. And we

will take you to dinner. I will say, but but in all seriousness, like listen, do I do I and some of the other execs with me struggle a little bit with should we raise to to do a big marketing push? And

I always go back to it's a little the Walmart idea. It's like, you know what,

Walmart idea. It's like, you know what, I want to be so thin on the margin. I

want them to know that their dollars are going to support the platform and that's where it's going and it's working.

People appreciate the people that are using it are doing the marketing for us.

They're telling their friends, "Hey, you see this thing? This thing is actually pretty good to use."

>> So, you see yourself a little bit like a plumber building pipes. And you say pipes don't change. Is that like in a zen way or we gave up on innovation?

>> Well, the the the auction we're talking about the best methods of yield to sell something. So, it's I'm not an academic.

something. So, it's I'm not an academic.

Wasn't the strongest part of my life history, but I'm pretty sure that the more buyers for something, the better off and more likelihood it is to sell

for a decent price. And that auction process to the most possible buyers.

That is exactly what the current exchangebased technology is built on. or

RTB specs is this idea that we're going to put this ad up for auction and you can build a tool that can see all the relevant information of the ad and submit a bid. AI and innovation and

things is going to happen all over the place here everywhere. But even if you ask AI what is the most efficient way to do this, it's going to come down to an auction system that's instantaneous

where there's as many buyers as possible and they all have as much information to make the decision and they can submit bids. It's not I just don't think the

bids. It's not I just don't think the core SP auctioning to DSP RTB is going to change. I I don't think that there's

to change. I I don't think that there's anywhere for it to improve in a way. I

mean again not not >> well I mean some people are getting rid of SSBs. Isn't that a big change?

of SSBs. Isn't that a big change?

>> That is from the last 10 years.

>> That's moving the auction. It's moving

the auction but there's still an auction but it's moving the auction a little bit more into the ad server than it is sitting in the SSP. No, not it's true.

That's just kind of that's definitely a change and there will be a lot of changes, but in terms of like the actual ad passing of information. I I I that's a great example you made, but I don't see many other examples that are going to and that doesn't actually change the

system. It just kind of moved it from

system. It just kind of moved it from one group to another. I think and I could be wrong. I'm I'm wrong all the time, but I think that that's not going to change. I don't I don't That's my

to change. I don't I don't That's my gut.

>> So AI hype is everywhere. Are you

secretly laughing at your competitors trying to sell AI powered faucets? I

mean we have built a full API into Pontiac. So many of our clients have

Pontiac. So many of our clients have built really slick AI style buying tools on top and they the prompt the prompt DSP and and others have built great

tools for other specific verticals and other specific things that advertisers care about. We have I think like a good

care about. We have I think like a good DSP should we have a bunch of clients who are doing very smart things that are that are building their own platform. So

what we see is brilliant innovation around the AI DSPs and the interfaces and all of that and we support it. Our

specific piece of technology is a little bit more just about managing this programmatic auction not you know my team wants to build some more AI stuff and we do some nice natural language

processing stuff for contextual and you know again we've but the core bit or the core technology for allocation of dollars by one by one that's that's where we spend most of our time and and

resources improving and that's and that's uh and again I I think that these AI DSPs are cool we're we're trying We're this isn't Andrew Casale from index line, but we're trying to be the

pipes. We believe that uh that will go

pipes. We believe that uh that will go through these pipes. We believe that it will be a low margin position when it is all said and done. And again, how big it gets is obviously my mission, but I

don't it's hard to say. But I do think that we we can earn a spot there in terms of the frontends and all the AI tools and all that. I think that will power some of them and I'm sure we will incorporate it. But no, it's not we're

incorporate it. But no, it's not we're not we're just not short-term people.

We're building a tool that we want to make sure that hundreds of companies can use to rely on managing their ad dollars and understand that that's going to take some time and and flashy things that are

that are really appetizing for for somebody today may get you in and you and you get booted to the curb 6 months, 12 months later. Again, we're we're uh we're trying to play the long game, get

into clients for many years.

>> So, if Pontek was a bathroom fixture, are you the shower, the sink, or the bedet? And why? All right,

bedet? And why? All right,

>> I almost was throwing a colonic in there, too. But

there, too. But >> that is an amazing one. Okay, bet. Bedet

is two is too >> cuz we know adtech needs a good cleanse.

>> I mean, listen, like I guess we'll go bed. And the re and the reason is is

bed. And the re and the reason is is because like it just could always be cleaner, you know, and like I think that I think that our focus on the publishers

>> like the char of tech, >> the focus on publishers, we want to talk about what websites you're on. We want

to talk about what pages of what websites, what content were they watching, what podcast were they listening to, did your user listen to a podcast and then watch a CTV show, which

shows, which are they listening to, what ads worked on display? like just by the thin margin thing and the allocation conversation and the promotion of PMPs

which is publisher market publisher like that helps you know get rid of some of the stuff third party data and we have by the way the ability to use data and we're plugged into the TransUnion marketplace and like we you know again

it's not like we had to we still have to be able to run campaigns if people want to run them but like you know in these in these DSPs people had 30 audiences from 16 different companies it's just

it's just too much. It could be cleaner and and I think that we help uh I think we help the user feel just extra clean and that's why we'll take the bed over.

>> Okay, we're in the time of a show where we we wreck you on a shipwreck you on a island, not wreck you personally and you mentioned that it's you your poor family you're bringing with you and Jack

Johnson, how did you survive? Is this

Naked and Afraid DSP edition or Gilligans Island with CPM math? I mean,

if you're going to say bootstrap and you're going to you're going to make it work and you're gonna you're going to work with all the resources you're given to make sure that you are efficient and

and succeed, I believe it. I believe I'd have to go naked and afraid route.

>> So, Jack Johnson is the singer.

>> Jack Johnson. Yes. The singer

songwriter. Music is great for the beach. You know, I just think that he's

beach. You know, I just think that he's he's uh I don't know exactly how old he is, but could he could help build some shelter?

He'd be able to help.

>> How useful is he? Can he make pancakes?

>> I mean, he's a beach guy, so I'm not I'm not like a total >> banana pancakes, maybe.

>> Yeah, I'm assuming the island is small like a cartoon. Maybe it's big. I don't

know. Is it a night? Is it like a abandoned island? I don't know. But I

abandoned island? I don't know. But I

was thinking that if I'm going to be >> It's abandoned. It's not like, you know, [snorts] Maui.

>> I think he lives on Maui. But uh but yeah, the uh yeah, that was my that was my he does [laughter] entertaining.

>> That's a coincidence. So your kids build a farm. Which crap crop do you plant

a farm. Which crap crop do you plant first? Corn, coconuts, or weed? And only

first? Corn, coconuts, or weed? And only

two can make you wealthy, obviously.

>> Am I on the island or is this now? I'm

not on the island. This is normal society.

>> No, this says you're still on the island.

>> I'm still on the [laughter] island. Corn

and weed.

>> Corn and weed.

>> Yeah, Jack Johnson is going to be very insistent on the weed. I

>> He's a big weed guy.

>> Yeah. And he's I don't want >> I mean it is Maui.

>> We're stuck. We're stuck on this island together. I want him to be comfortable

together. I want him to be comfortable and be able to you know >> you can make a bong out of a coconut.

>> He's going to he'll handle all.

>> So how would you explain would you rather explain programmatic auctions to a coconut or try to teach your kids to clean up after themselves?

>> I've made a you know coconut you know like like a Wilson type thing from Castaway. I mean, that's going to be

Castaway. I mean, that's going to be >> it's really >> I'm sure that reference is gone for some people. Like, they just have no idea

people. Like, they just have no idea what we're talking about. I think I brought up Wilson to like a Gen Z person and they just looked at me like I had lost my mind. Like, you know, person was

on an island talking to a Coke and they just like backed away. Okay, cool man.

>> I mean, I've made a career explaining programmatic to people. That is that is if you ask me what I've done, it's I I >> got to ask you then. If you had explained programmatic to your kids, and

maybe you have, how did you do it?

>> Around this instantaneous auction thing, around this idea that, you know, they're getting a little bit older conveniently, and they watch themselves, uh, Curious George streaming and some other things,

and there are ad slots, and there's ads that fire there's ads that fire two in the pod, and there's ads that have there's there's there's on CTV there's often ad slots with no ads in them.

Commercial playing, commercial airing.

So, >> and you're trying to sell it to somebody.

>> Yeah. Well, yeah. I've I've to my kids.

I explained it piece by piece by saying, "Listen, there's tools that are trying to sell this in real time and they either didn't find a buyer or no tool bid on it." And that's one of the tools

that we've built. Your dad and Eric and the people that they've met a few times.

And uh and I explain that it's all based on this like really quick auction. And

it happens every time they see an ad on TV. It happens every time they see a

TV. It happens every time they see a digital billboard. Happens. doesn't

digital billboard. Happens. doesn't

happen every time they see [laughter] I have I have Netflix I I see those no ads all the time.

>> Yeah. No,

>> I'm just staring at the screen waiting.

And the funny thing is I'll be staring at the screen for like five minutes and then my wife and I cuz we don't actually have a TV so we have the Kindle and we're just like, you know, we've been staring at a blank screen for five minutes and we have no idea. Like we

just let it go.

>> Not they're not sold. None of those CTV guys are sold out. I probably don't want to say that but that's that's the truth.

So, let's just say Jeff Green comes in his um large yacht and tries to rescue you. First of all, would you take the

you. First of all, would you take the rescue from Jeff Green? And second of all, what's the first ad campaign you'd launch to tell the world how you survived?

>> I mean, lowering the margin in any business is very difficult. Jeff has

been very big proponent of focusing all of his attention and the TTD's attention on the product that they've built. They

are 100% focused on the >> Yeah. At this point, I don't think he's

>> Yeah. At this point, I don't think he's going to rescue you.

>> I Well, well, hold on. I I This is very unlikely, but And I'm not drowning. I'm

I'm we're we're we're partying on our boat, but uh >> Yeah, maybe he's coming to visit you.

Maybe it's the new >> Well, maybe my boat could like slip up to the side of his I don't know. This

the analogy is tough, but >> you're probably seen as a terrorist. Is

there a world that our DSP starts managing kind of again more of the allocation side and it's at a margin that's you know a third of theirs or whatever and there's there's it's a

different business model. It's a

different business model. Listen

obviously they would have to change from their we only run TTD's platform but if he gets into the business of buying other advertising platforms could hedge him against the low margin problem maybe

there would be some type of fit.

unlikely. Very unlikely, but I wouldn't, you know, never say never.

>> I'm not sure that was an answer to the question, but we'll go with it. So,

about a little bit about you. You're the

kid from Union College with a degree in mechanical engineering.

>> Yes, sir.

>> You have three kids, a wife, a background in numbers, and you love advertising as you see it enduring in culture. You said it's really hard, but

culture. You said it's really hard, but you never leave. What is the reason? And

how is it Stockholm syndrome? I mean, a job making money, you know, it's not easy. There's a lot of very difficult,

easy. There's a lot of very difficult, very hard, very high pressure jobs. Not

that >> What's the hardest part? The people or or what?

>> Getting in is very difficult. This is a small community. You know that it's not

small community. You know that it's not that many people in this business, especially if you focus just on the media side, it gets smaller. So, it's

hard to get in. It's hard to learn how to use these platforms. You can't really learn how to use a DSP unless you manage clients budgets. Why is a client giving

clients budgets. Why is a client giving you a $100,000 as a budget for you to learn on? I don't think that's going to

learn on? I don't think that's going to happen. So, there's a chicken and egg

happen. So, there's a chicken and egg issue of how you learn and how you get in.

>> Has anyone ever considered gifying it?

Like doing a DSP game?

>> Well, >> training.

>> Yeah. I mean there's trainings is one thing but training doesn't you know if you if you spend money a day that the campaign is not supposed to be live.

It's a feeling of stress that is impossible to teach and impossible to coach without going through it. As you

learn to fly a plane, there's periods of the training where they shut down the engines while you're in the sky. You had

to get in the plane to get the engines shut on you. That's similar to, you know, you can't uh you can't necessarily do that here. So it's hard to get in once you're in and you understand it. I

find it rewarding and exciting and and interesting and technology. We're on the edge of technology and how things work and computer speed and and innovation and I just think it's uh I think it's

got a lot of the components to keep it very interesting and very and fulfilling. Uh yeah, I wouldn't say this

fulfilling. Uh yeah, I wouldn't say this is never the case because there's campaigns that touch on very hard subjects and things that are very important and privacy, but generally it's not a life and death type business.

Generally, you're you're not dealing with that type of thing. So, I think advertising is is just a good place to be. It is, you know, it's competitive

be. It is, you know, it's competitive and it's tough and it can burn you out when they especially when you're young and they grind you at these agencies.

Like I I can understand people getting burned out, but I always recommend people take a change. They they change the side of the business, but I find that it's a very very good business.

I've seen a lot of people leave and come back. I'm sure you have too.

back. I'm sure you have too.

>> So, who is your Yoda and what do they teach you? And for shits and giggles,

teach you? And for shits and giggles, you have to say it like Yoda.

>> So Paul Roscowski is my partner. And

Paul came in. I was at Varic Media Management, which was the trading desk arm of MDC Partners, which is now kind of morphed into Stagwell through acquisitions and whatnot. When I was there, I was a I was a trader. I was a

media trader. And the management came in

media trader. And the management came in and said, "They're hiring a new president." And they brought on Paul

president." And they brought on Paul Roscowski, who's now my for years, I called him my mentor. Now I call him my partner. Paul came uh with an attitude

partner. Paul came uh with an attitude on how to drive sales and innovation that to me showed me that while you had

to make money and you got to close deals and you got to perform on the media campaigns and you got to do the job, he really showed me how it could be fun. I

mean, he he showed me how you travel to Dallas for a meeting and you catch another flight and you're in Kansas City and then you're at dinner in Minneapolis and you're you're pushing relationships,

but ultimately with with a little bit of the idea that it it it can be it can be fun. There's no rules to how you build a

fun. There's no rules to how you build a successful business. There's there's

successful business. There's there's some rules obviously, but there's really no it's not about doing it the way somebody else did it. It's not about doing it a certain way. as long as we're

able to do it legally, fairly, and with a with in good nature, there are almost no rules. He and I got along very

no rules. He and I got along very quickly, and he promoted me from just a trader to the head of the trading organization. But I guess to make this

organization. But I guess to make this uh to to close this segment full circle and give you a little insight into Paul, he thought that everybody in the organization just kept coming to me with questions. They just kept coming to me.

questions. They just kept coming to me.

Just go to Keith. How does that work? Go

to Keith. What's the expected close rate on that campaign? Go to Keith. Go to

Keith. Go to Keith. Wasn't a very big company. I mean, it was we're talking

company. I mean, it was we're talking about a 15 person company that got up to about 50 before I left. But for a period there, and this changed as I grew out of the director role, but he thought it was

funny to call me the director of yield operations, data, and analytics. And uh

that was the director of Yoda as for short.

>> So, what's the Yoda phrase? Like fun

media buyin is >> insure. [laughter] Yeah, I know. It's

>> insure. [laughter] Yeah, I know. It's

hard putting you on the spot.

>> Clean. As clean as can be. This clean.

No, I don't know. Hard to say.

>> Superpower time. You said you wanted speed of light travel, which is actually probably the smartest thing I've heard because you can also do time travel with time dilation. If you had it, what's the

time dilation. If you had it, what's the first double book trip you'd fix, >> man? I mean, clients,

>> man? I mean, clients, >> we going back to Maui.

>> Well, yeah. And also hanging with Jack on our island and doing I would be able to do both. That'd be nice. So, what's

the craziest industry opinion that make people look at you like you have lost your mind? Like probably cookies are

your mind? Like probably cookies are bad.

>> I mean, yeah, the I would say if I walk into a room and I'm like, "Hey, you're going to get better performance on Pontiac than the DSP you're using today." I think most people look at me

today." I think most people look at me like I'm insane. But a DSP is not a DSP is a way into an exchange. I mean it it's it's not like we talk about this in

the financial uh in the financial the financial comparisons of the exchanges have been have been done for years and it doesn't do >> I mean that was the that was the inspiration originally.

>> Exactly. But like in all seriousness um we're talking about a tool that's used in the finance world to buy and sell stocks. Fidelity trade

sell stocks. Fidelity trade a lot of tools lot of tools to do it.

>> Do they actually do the performance any of these? No.

They just give you a way to buy and sell them and and and you have to do the performance, but you want a tool that works good viably all the time, sees all the stocks, sees all the in our space, ad space in our in our case. So, it's

not the platform, it's it's the what you're doing. And obviously, a platform

you're doing. And obviously, a platform can encourage you to do it better, but especially in that space, conveniently, low fees. That's the first thing people

low fees. That's the first thing people promote, low fees. Why do they promote that? Because less of the money is going

that? Because less of the money is going to the middle, more of it's going into the stock. This is the same game. We're

the stock. This is the same game. We're

We're good at performance. Did we build a good algorithm? Does the tool do what you want? Oh, yeah. Does it also clear

you want? Oh, yeah. Does it also clear the ads for a lower price because the margin's lower? Yes.

margin's lower? Yes.

>> So, looking back in 20 years, what do you want your legacy to be? The numbers

guy who built DSPs or the dad who got your kids to clean their room?

>> I mean, I manage an organization of 35 folks. I take managing people and their

folks. I take managing people and their career choices and their work very seriously. I'm proud of the team that I

seriously. I'm proud of the team that I that I have at the moment and people that have come and gone. And I I take that very seriously. I think that it's it's important to build DSPs. It's

important to make a value added to or service that actually deserves to be paid and make money. It's okay to make money. We're not a nonprofit. But I take

money. We're not a nonprofit. But I take the human component and the way that the company treats its employees and people in the circle very very seriously. And

that's I think that that if I had to hope I think that the legacy would be people would say that they liked working at the company. They thought it was a good mission. Hopefully they made some

good mission. Hopefully they made some money and that that would mean a lot more to me having a having a a group however big. Hopefully it grows to

however big. Hopefully it grows to hundreds of people. But I would I would say I hope the my kids clean and the people that I've worked with closely think of it.

>> That's a wrap for today's episode of the Adat show. CTV is blowing up faster than

Adat show. CTV is blowing up faster than my inbox after I say something mildly honest on LinkedIn. And Keith Gooberman just gave us the kind of DSP [music] reality check the industry usually needs

a stiff drink for. A quick thank you to the people who keep his circus running.

Troutman Amin LLP, the legal sherpers who somehow managed to keep half the adtech world from tripping over its own [music] shoelaces. and incremental.com

[music] shoelaces. and incremental.com proving that actual measurement isn't a mythical [music] creature we whisper about in conference hallways. Stay bold,

stay curious, and know more than you did yesterday.